Pergamon Bioorganic & Medicinal Chemistry Letters, Vol. 7, No. 5, pp. 581-586, 1997

© 1997 Elsevier Science Ltd All rights reserved. Printed in Great Britain

0960-894X/97 $17.00 + 0.00 PII: S0960-894X(97)00048-6

CHARACTERIZATION OF CYCLIC ADP-RIBOSE AND 2'-PHOSPHO-CYCLIC-ADP-RIBOSE BY 31p NMR SPECTROSCOPY.

Klaus D. Schnackerz*, Chinh Q. Vu #, David Gani $, Rafael Alvarez-Gonzalez +, and Myron K. Jacobson # Theodor-Boveri-Institut f#r Biowissenschaften, Physiologische Chemie I, Univers'itgit Wiirzburg, Am Hubland, D-97074 W#rzburg, Germany, #Division of Medicinal Chemistry and Pharmaceutics, College of Pharmacy, and Lucille P. Markley Cancer Center, Universi(y of Kentucky, Lexington, KY 40536-0082, $School of Chemistry, University of St. Andrews, St. Andrews, Fife, KYI 6 9S~ UK, and +Dept. of Micro- biology and Immunology, (hfiversi O, of North Texas Health Science Center, Fort Worth, TX 76107, U.S.A.

Abstract: The 31p NMR spectra of the Ca 2+ mobilizing nucleotides cyclic ADP-ribose and 2'-phospho- cyclic ADP-ribose have been studied. Both nucleotides show doublets of doublets in the range -9.5 to -11 ppm for the pyrophosphate group. The P-cADPR exhibits a singlet in the 0 to +4 ppm range for the 2'- phospho group. The pH dependence of the 31p signals shows transitions with pK a values of 8.37 and 8.95 for cADPR and P-cADPR, respectively. © 1997 Elsevier Science Ltd. All rights reserved.

Introduction

Cyclic ADP-ribose (cADPR), a metabolite of NAD, is a potent calcium releasing agent which is

believed to be a second messenger involved in the regulation of calcium homeostasis.l-3 The synthesis of

cADPR from NAD involves the cleavage of the nicotinamide from the ribose and cyclization of the ribose

to the N-1 of the adenine moiety4, 5 (Fig. 1). In mammalian cells, cADPR is synthesized and degraded by a

multifunctional enzyme that contains cADPR synthase, cADPR hydrolase, and NAD glycohydrolase

activities. 4 More recently, an analogous cyclic nucleotide synthesized from NADP, 2'-phospho-cyclic

ADP-ribose (P-cADPR) (Fig. 1), has been shown to elicit Ca 2+ release from rat brain microsomes. 6,7 Tile

characteristics of Ca 2+ release indicate that P-cADPR and cADPR elicit Ca ?+ release by a similar

mechanism, but by a mechanism distinct from inositol trisphosphate. 7 The presence of cADPR in animal

HO OH HO OH

0 a- ~- 0

"O-P-O-CHz N NH

' " " 2 ' - - , f ' T "O-P-O-CH2 !, / / , H* -O-~-O-C" , , ~ / / N

II \ -o "~- - f I I ' , " o " - - : '

HO OR HO OE

Figure 1. Possible structures for the acid and free base forms of cADPR and P-cADPR

R = H, cADPR; R= PO3 2, P-cADPR

* E-mail: schnacke@bJozentrum, unJ-wuerzburg, de. Fax: (+49) 931-8884150

581

582 K.D. SCHNACKERZ et al.

cells is well established 8 and the natural occurrence of P-cADPR has been recently detected.9

Characterizations of cADPR and P-cADPR by 1H-NMR and UV spectroscopy have provided

information concerning the configuration of these nucleotides.7,1°-12 While cADPR and P-cADPR are

potent calcium mobilizing agents, ADPR and P-ADPR are totally inactive, 7 indicating that the cyclic

nucleotides adopt a unique conformation(s) that results in the activation of membrane Ca 2+ channels. We

report here the 31p NMR spectra of both cyclic nucleotides as a function of pH. Our results indicate that

the 31p signals of these nucleotides show a pH dependence which correlates to spectral changes in uv

absorbanee for the adenine chromophore. Such protonation/deprotonation events may be important in

calcium channel activation.

Materials and Methods

Chemicals. cADPR and P-cADPR were synthesized and purified as described previously. 4,7 2H20

was purchased from Sigma. All other chemicals were of highest purity commercially available.

31p NMR Spectroscopy. Fourier transformed 31p NMR spectra were collected at 121.497 MHz on a

Bruker AM300 SWB superconducting spectrometer using a 10-ram multinuclear probehead with

broadband IH decoupling The NMR tube spinning at 15 Hz contained the sample (2 mL) and 2H20 (0.2

mL) as field/frequency lock and was maintained at 20 + 0.1 °C using a thermostated continuous air flow.

Generally, a spectral width of 3000 Hz was acquired in 8K data points with a pulse angle of 60 °. The

exponential line broadening prior to Fourier transformation was 1 Hz. The samples were dissolved in 50

mM Hepes buffer and the pH was adjusted by addition of acid or base. pH values of the samples were

determined before and after the NMR measurement. Positive chemical shifts in ppm are downfield changes

with respect an external chemical shift reference standard of 85% phosphoric acid.

Results and Discussion

The { 1H}31p NMR spectrum of cADPR at pH 7.06 shows the expected doublet of doublets in the

range -10 to -t 1 ppm for the pyrophosphate group (Fig. 2A). The similar location of the two doublets

indicates that both phosphate groups are substituted. The singly substituted pyrophosphate groups of

Characterization of cyclic ADP-ribose 583

1t , , i i

-'g -[0 -h -[2 -[3 PPH

Figure 2: 31p NMR spectra of cADPR (A) and P-cADPR (B) at pH 7.06 and 20°C. Conditions: 5,800 scans (A), 4,300 scans (B); IH broadband decoupling, line broadening prior to Fourier transformation 1 Hz Inset in (B) shows the signal of the 2'-phospho group of P-cADPR

thiamine diphosphate 13 and ADP 14 are separated by 4 and 4.6 ppm, respectively. The phosphorus-

phosphorus spin coupling (Jal3) in cADPR is constant at 136 Hz over a range of pH values (6.0-10.0),

whereas the coupling constants of FAD and ADP are 20.9 and 22.0 Hz, respectively. 15 Note that in a

control experiment the spin coupling constant for ADP-ribose was determined to be 21.9 Hz (data not

shown). The smaller spin coupling constant for cADPR most probably results from the conformational

restrictions imposed upon the pyrophosphate group by cyclization. The coupling constant for cADPR

obse~wed in this study closely matches the value of 14.5 Hz reported by Wada et al. 12 at unspecified pH.

Fig. 3 shows the 31p chemical shifts of cADPR for the pH range 6.8 to 10.0. One of the two phosphate

moieties occurs at -10.2 ppm and shows very little change as a function of pH, while the second exhibits a

strong pH dependence. The transition for this latter phosphate group shows a pK a value of 8.30 and

limiting 8 values of-10.83, and -9.86 ppm at low and high pH values, respectively. Wada et al 12 reported

3 l p chemical shifts of-9.92 and -10.67 ppm in deuterium oxide for cADPR at unspecified pH and assigned

the former to the adenosine unit and the latter to the ribose unit attached to N-1 using 1H-31p COSY

experiments. We believe that these assignments for the 31p NMR signals may not be correct and should be

reversed. The 1H-31p and 1H-IH coupling constants for the 4' and 5' protons in each ribose furanosyl

moiety are difficult to determine from the reported spectra and some are similar. Moreover, the symmetry

of the cADPR about the pyrophosphate O atom, together with its cyclic structure, make comparative

584 K.D. SCHNACKERZ et al.

assignments difficult. It has been shown that for noncyclic mono- and disubstituted adenosine

diphosphates, the 31p signal for the phosphate group closest to the adenosine unit, the (x-phosphate,

occurs at a higher field than that for the J3-phosphate group. 13-15

In contrast to cADPR, the 31p NMR signals for each of the two phosphate groups of ADP-ribose

occur as doublets at -10.47 and -10.64 ppm, respectively (data not shown) and the chemical shifts of these

signals are pH-independent over the range 6.4 - 9.4. At pH values above pH 7.0, the distance between the

doublets of cADPR diminishes until approximately pH 8.6, at which point a single signal remains. Above

pH 8.6, the signals cross over and the distance between them increases.

The 31p NMR spectrum of P-cADPR at pH 7.0 exhibits a singlet at 3.74 ppm and a doublet of

doublets at -9.93, and -10.92 ppm (Fig. 2B). The singlet is assigned to the 2'-phosphomonoester group

(Fig. 1). The pH dependence of the singlet shows a single transition with a pK a of 5.89 and limiting 6

values at -0.05 and 4.15 ppm for the low and high pH values, respectively (data not shown). The pK a

value is somewhat lower than those of normal phosphomonoesters. 16 The pH dependence of the doublets

E

&

-10

- l l

0 0 0 0 0 0

• 0

l , I .. , I i I

7 8 9 10 pH

Figure 3. pH dependence of the 31p chemical shift of cADPR and P-cADPR . The solid lines were derived by iterative computer analysis. 13 For cADPR (D, I ) a pK value of 8.30 and limiting 6 values of -9.85 and -10.83 at low and high pH values, respectively and for P-cADPR (O, O) a pK value of 8.96 and limiting 6 values of-9.83 and -10.86 were calculated.

Characterization of cyclic ADP-ribose 585

for P-cADPR in the pH range 6.8 - 10.0 shows that the phosphate moiety at -9.95 ppm is insensitive to pH

changes while the other phosphate shows a single transition (Fig. 3). For this doublet, a pK a value of 8.91

and limiting 8 values of-10.87 and -9.88 ppm for the low and high pH values, respectively, were obse~ed

Previous work has shown that the third and fourth pK values (pKa3 and PKa4 ) for inorganic

pyrophosphate are 6.25 and 8.64, respectively. 13 Since the PKal and PKa2 values of a disubstituted

pyrophosphate group are both below 2.0,13 the observed pK a value of 8.30 for cADPR must correspond

to the protonation/deprotonation of the adenine moiety. In principle, deprotonation of the adenine ring

system could cause a change in the chemical shift of either one or both of the 31p signals. However, only

one of the signals in both cADPR and P-cADPR is affected by the titration and, indeed, in each case it is

for the same phosphate group. While anisotropic effects cannot be excluded, the magnitude of the change

and the fact that completely analogous pH-dependent modifications occur in the UV spectra suggest the

possibility of a direct interaction between the reporter phosphate group and the titrated chromophore A

possible structure for the acid forms would be the N-3 protonated vinylogous amidine of the adenine

system which displays characteristic )~max values at around 270 nm. 17,t8 Any such N-3 protonated

chromophore could form a hydrogen bond with oxygen atoms &the adenosine phosphate moiety and in so

doing would provide a mechansim for transferring electronic information from the chromophore to the P-

atom Alternatively, when the protonation/deprotonation occurs distant from the phosphorus nuclei,

conformational restrictions imposed upon the pyrophosphate group by cyclization could be sensed by the

phosphorus nuclei since the 31p chemical shift of phosphate esters correlate with small changes iv, O-P-O

bond angles. 16.19

Thus, we postulate that for cADPR and P-cADPR the signal which displays a pH-dependent

chemical shift is that for the phosphate group that is nearest to the ribose attached at the N-9 position of

the adenine moiety.

The spectral data presented here may prove useful for future studies aimed at understanding the

molecular basis of Ca 2+ channel activation by cADPR and P-cADPR The smaller phosphorus-phosphorus

spin coupling constants of cADPR and P-cADPR compared to the noncyclized forms of these nucleotides

and other dinucleotides indicate that cADPR and P-cADPR have a solution structure that has

conformational restrictions for the pyrophosphate moiety. This conformation may be related to the ability

of these nucleotides to interact with a Ca 2+ channel protein to achieve activation. Additionally, the spectral

changes observed upon protonation/deprotonation may be related to the mechanism of channel activation

586 K.D. SCHNACKERZ et al.

as proton abstraction/addition to the protein-bound nucleotide may initiate a conformational change in the

nucleotide that in turn translates into channel activation. The spectral differences between protonated and

deprotonated forms should also be useful for the study of the interaction of these nucleotides with their

binding proteins once these proteins have been identified.

Acknowledgment. This study was supported in part by NIH Grant CA43894 to MKJ and GM45451 to

RAG and by a grant of the Deutsche Forschungsgemeinschaft to KDS. We thank Biotechnology and

Biological Sciences Research Council for an International Scientific Interchange Scheme Grant to

Wurzburg for DG. We are indebted to Donna Coyle for excellent technical support and to Dr. Manfred

Buehner for the computer analysis of the pH dependence &the chemical shifts.

References and Notes

1. Koshiyama, H.; Lee, H. C.; Tashajian, A. H. J. Biol. Chem. 1991, 266, 16985-16988. 2. Takasawa, S.; Nata, K.; Yonekura, H.; Okamoto, H. Science 1993, 259, 370-373. 3. Lee, H. C.; Galione, A.; Walseth, F. F. VitaminHormon. 1994, 48, 199-257. 4. Kim, H.; Jacobson, E L.; Jacobson, M. K. Science 1993, 261, 1330-1333. 5. Lee, H. C.; Aarhus, R.; Levitt, D. Nature Structure Biology 1994, 1, 143-144. 6. Zhang, F.-J.; Gu, Q.-M.; Jing, P.; Shih, C.J. Bioorg. Med. Chem. Lett. 1995, 2267-2272. 7. Vu, C. Q.; Lu, P. J.; Chen, C-S.; Jacobson, M. K. J. Biol. Chem. 1996, 271, 4747-4754. 8. Walseth, T. F.; Aarhus, R.; Zeleznikar, R. J.; Lee, H. C. Biochim. Biophys. Acta 1991, 1094, 113-120. 9. Vu, C. Q.; Coyle, D. C.; Jacobson, M. K. unpublished data. 10. Kim, H.; Jacobson, E. L.; Jacobson, M. K. Biochem. Biophys. Res. Commun.1993, 194, 1143-1147. 11. Gu, Q-M.; Sih, C. J. JAm.. Chem. Soc. 1994, 116, 7481-7486. 12. Wada, T.; Inageda, K.; Aritomo, K.; Tokita, K.; Nishina, H.; Takahashi, K.; Katada, T.; Sekine, M

Nucleosides & Nucleotides 1995, 14, 1301-1314. 13. Lachmann, H.; Schnackerz, K. D. Org. Mag7t. Resort. 1984, 22, 101-105. 14. Vogel, H J.; Bridger, W. A. Biochemistry 1982, 21,394-401. 15. Kainosho, M.; Kyogoku, Y. Biochemistry 1972, 11,741-752. 16. Gorenstein, D. G. J. Am. Chem. Soc. 1975, 97, 898-900. 17. Specker, H; Gawrosch, H.; Chem. Ber. 1942, 75, 1338-1338. 18. Brookes, P.; Lawley, PD.; J. Chem. Soc 1960, 539-545. 19. Schnackerz, K. D.; Wahler, G.; Vincent, M. G; Jansonius, J. N. Eur. J. Biochem. 1989,185, 525-531.

(Received in Belgium 20 November 1996; accepted 24 January 1997)